Model and experiments to optimize co-adaptation in a simplified myoelectric control system.

OBJECTIVE: To compensate for a limb lost in an amputation, myoelectric prostheses use surface electromyography (EMG) from the remaining muscles to control the prosthesis. Despite considerable progress, myoelectric controls remain markedly different from the way we normally control movements, and require intense user adaptation. To overcome this, our goal is to explore concurrent machine co-adaptation techniques that are developed in the field of brain-machine interface, and that are beginning to be used in myoelectric controls. APPROACH: We combined a simplified myoelectric control with a perturbation for which human adaptation is well characterized and modeled, in order to explore co-adaptation settings in a principled manner. RESULTS: First, we reproduced results obtained in a classical visuomotor rotation paradigm in our simplified myoelectric context, where we rotate the muscle pulling vectors used to reconstruct wrist force from EMG. Then, a model of human adaptation in response to directional error was used to simulate various co-adaptation settings, where perturbations and machine co-adaptation are both applied on muscle pulling vectors. These simulations established that a relatively low gain of machine co-adaptation that minimizes final errors generates slow and incomplete adaptation, while higher gains increase adaptation rate but also errors by amplifying noise. After experimental verification on real subjects, we tested a variable gain that cumulates the advantages of both, and implemented it with directionally tuned neurons similar to those used to model human adaptation. This enables machine co-adaptation to locally improve myoelectric control, and to absorb more challenging perturbations. SIGNIFICANCE: The simplified context used here enabled to explore co-adaptation settings in both simulations and experiments, and to raise important considerations such as the need for a variable gain encoded locally. The benefits and limits of extending this approach to more complex and functional myoelectric contexts are discussed.

Temporal information for spatially constrained locomotion.

How is locomotion adapted to spatial environmental constraints? The control of this everyday behavior is claimed to be based on information that specifies either spatial or temporal properties of the actor-environment system. Although studies on open-loop locomotor pointing (i.e., the positioning of a foot on a target on the floor while walking) agree on the use of spatial information, studies on closed-loop locomotor pointing propose the use of temporal information. Here, we test the hypothesis of closed-loop locomotor pointing based on temporal information, by dissociating spatial and temporal information in a virtual reality setup (virtual environment connected to a treadmill). The results support this hypothesis and shed some light on the type of temporal information that is used. The performed dissociation between spatial and temporal information, however, does not rule out a control based on a continuous updating of spatial information. Therefore, our conclusion on the use of temporal information was moderated.

Perception--action coupling model for human locomotor pointing.

How do humans achieve the precise positioning of the feet during walking, for example, to reach the first step of a stairway? We addressed this question at the visuomotor integration level. Based on the optical specification of the required adaptation, a dynamical system model of the visuomotor control of human locomotor pointing was devised for the positioning of a foot on a visible target on the floor during walking. Visuomotor integration consists of directly linking optical information to a motor command that specifically modulates step length in accordance with the ongoing dynamics of locomotor pattern generation. The adaptation of locomotion emerges from a perception-action coupling type of control based on temporal information rather than on feedforward planning of movements. The proposed model reproduces experimental results obtained for human locomotor pointing.

Spatially constrained locomotion under informational conflict.

This study investigates the informational based that supports intentional adaptation of locomotion to spatial environmental constraints. A virtual reality setup was used to present subjects with targets providing normal as well as abnormal optical expansion during locomotor pointing (i.e. positioning of a foot on a visible target on the floor during walking). The manipulation dissociated two variables providing temporal information about time-to-passage (TTP): TTP(beta alpha) which encompasses target expansion, and TTP(alpha) which is independent of target expansion. While a previous study showed TTP(alpha) to be sufficient, the present results reveal that TTP(beta alpha) may be used when it is available. This finding indicates that both variables play a role that varies according to the circumstances. Furthermore, the present results provide evidence of the operation of a security principle for action in conflicting situations.

The control of human locomotor pointing under restricted informational conditions.

Since a perception-action coupling type of control (Kugler, P.N. and Turvey, M.T., Information, Natural Law, and Self-Assembly of Rhythmic Movements, Erlbaum, Hillsdale, 1987, 481 pp.) continuously operates during locomotor pointing tasks (e.g. long jumping) (Montagne, G., Cornus, S., Glize, D., Quaine, F. and Laurent, M., A 'perception-action coupling' type of control in long-jumping. J. Motor Behav., (2000) in press), the information sources underlying this control have to be dealt with. Under the assumption that subjects use information about the first-order time remaining before they pass the target, we identify in the literature four different sources of information that specify this physical property. Only one of these sources is inevitably present under all possible environmental conditions containing at least a continuously visible target on the floor. This study aimed to test its sufficiency to perform a locomotor pointing task. The use of a virtual reality set-up permitted us to compare locomotor pointing executed with all four information sources or only with the aforementioned one. The likeness found between those two conditions, as far as the pointing performance and the mode of control are concerned, expresses the evoked sufficiency.

The study of locomotor pointing in virtual reality: the validation of a test set-up.

The goal of this experiment was to validate an experimental set-up for studying locomotor pointing. The specific and also original element of this set-up was the interactive nature of virtual reality and movement production. This interaction was achieved through the coupling of a treadmill and a Silicon Graphics system. This latter system generated on a screen (3 x 2.3 m) an environmental array that moved according to the action produced by subjects on a treadmill. The task was to place either foot on a spatial target that appeared on the floor in front of the subject's displacement trajectory. We analyzed the step length patterns of subjects approaching these targets, along with the current target-subject relationship. The results are in agreement with a perception-action coupling type of control mechanism that operates continuously as the subject approaches the desired target. Apparently, these findings mirror observations of real-life locomotion, indicating that the present set-up provides a valid and useful tool for examining human locomotion.

The regulation of externally paced human locomotion in virtual reality.

The goal of the present experiment was to study regulation of human locomotion under externally paced temporal constraints. On a screen placed in front of a treadmill a virtual hallway was projected (Silicon Graphics systems) in which a pair of doors were presented that continuously opened and closed at a rate of 1 Hz. Subjects were attached to a locometer and instructed to regulate walking pace such that the doors were passed correctly. Performance outcome, movement kinematics (stride duration, stride length and synchronization of stride and door cycles) and flow patterns (change in visual angle of door aperture) were used to examine the data. The analysis of the synchronization patterns indicates that stride cycles were not linked to the period of door oscillation. Moreover, results for stride duration reveal that subjects walked at their preferred speed up to the final phase of the approach. This observation is supported by the inspection of the flow patterns, revealing a final increase in variability as a result of regulation. In sum, regulation of locomotion under externally paced temporal constraints seems to generate a specific functional behavior. It appears that regulations are postponed until the final stage of the approach during which adaptations are made according to the requirements of the current situation.